Hundreds of Mundane AI Apps Are Hiding Sexually Abusive Deepfake Features
Hundreds of AI apps disguised as benign tools secretly generate abusive deepfake content.
A WIRED investigation uncovered over 400 apps and websites, often developed by Chinese entities, that masquerade as innocuous services like pet portrait generators or AI workout coaches. However, these applications secretly load interfaces to create explicit AI-generated images and videos, including violent and sexually abusive content, when triggered under specific conditions. The network was promoted through tens of thousands of ads on Meta platforms, some containing child sexual abuse material, leading to investigations by global regulators.
AI summary of the linked source. Not reviewed by an editor. How this works
Run Meta?
Get an email when a new story names Meta, usually within the hour.
Free. Your list stays private and never appears in a subject line. One click stops it. How Vendor Watch worksPrivacy
The lesson from this incident
Full lessonOver 400 apps and websites concealed deeply harmful capabilities—including the generation of child sexual abuse material and violent deepfakes—beneath the surface of benign-looking AI services like pet portrait generators. These applications exploited the trust users place in mainstream platforms by advertising extensively on Meta, meaning the supply chain of app distribution became a vector for serious harm. The case highlights how malicious actors can weaponize AI tooling while hiding behind innocuous branding, making platform-level vetting and regulatory oversight critical. Without robust app review processes and transparent capability disclosure, end users have no reliable way to know what a downloaded application actually does.
What helps against this
How we list vendors
- SaaS security posture management (SSPM)
Business apps like Microsoft 365 and Salesforce stop leaking through risky settings and connected apps.
- AI application security
Prompt injection, data leaks through AI tools and unapproved AI use get caught.